Graduate school in Systems, Optimization,
Control and Networks (SOCN)
|
Graduate school courses scheduled in the academic year
2007-2008
"NUMERICAL METHODS FOR NONLINEAR OPTIMAL CONTROL
PROBLEMS"
Lecturer : Moritz DIEHL (Katholieke Universiteit Leuven, Belgium)
Spring 2008
"MECHATRONICS : DYNAMICS OF ELECTROMECHANICAL AND PIEZOELECTRIC SYSTEMS, WITH APPLICATIONS IN VIBRATION CONTROL"
Lecturer : André PREUMONT (Université Libre de Bruxelles, Belgium) CANCELLED"ROBUST TRACKING AND DISTURBANCE REJECTION "
Lecturer : Alberto ISIDORI (University of Roma, Italy)"CONSTRUCTIVE LMI APPROACHES FOR ANTI-WINDUP COMPENSATOR DESIGN FOR SYSTEMS SUBJECT TO SATURATION"
Lecturer : Sophie TARBOURIECH (LAAS-CNRS, Toulouse, France)
Detailed contents
1. OPTIMIZATION ALGORITHMS ON MATRIX MANIFOLDS
Lecturers : P.-A. ABSIL (Université Catholique de Louvain, Belgium) and R. SEPULCHRE (Université de Liège, Belgium)
Overview
The course will offer an introduction to the theory and applications of numerical
optimization on manifolds. This area of computational mathematics deals with
the analysis and design
of algorithms for optimizing a real-valued function whose domain, instead
of being the classical Euclidean space Rn, is in general a non-Euclidean space
endowed with a differentiable structure.
This includes real-valued functions on smooth surfaces of Rn, and more generally
on sets that can be “smoothly” parameterized. Optimization techniques
on manifolds find applications in several areas, including the eigenvalue
problem, blind source separation (via independent component analysis), pose
estimation in computer vision, and model reduction of dynamical systems.
• Motivate why manifold theory provides natural foundations for the development
of matrix algorithms for several major computational problems and why the resulting
numerical algorithms can be highly competitive.
• Provide the students with the necessary background in differential geometry
instrumental to algorithmic developments.
• Develop classical optimization algorithms on matrix manifolds: steepest-descent,
Newton, conjugate gradients, trust-region methods.
• Guide the student through the concrete calculations that turn an abstract
geometric algorithm into a numerical implementation.
• Illustrate the many problems from linear algebra, advanced signal processing,
and statistical analysis that can be recast as optimization problems on matrix
manifolds.
Contents
1. Motivating examples: Eigenvalue problem, independent component analysis,
computer vision, dynamical systems. Manifolds, submanifolds, quotient manifolds.
2. First-order geometry: Tangent vectors, Riemannian metric, gradient vector
fields, gradient flows.
3. First-order algorithms: Retractions, seepest-descent methods, line-search
strategies. Optimization on the orthogonal group and independent component analysis.
4. Second-order geometry and Newton’s method: Newton’s method in Rn,
affine connections, Riemannian connection, parallel translation, geodesics,
Newton’s method on manifolds.
5. Models and trust-region methods. Optimization of the Rayleigh quotient on
the sphere and on the Grassmann manifold.
6. Vector transport, approximate Newton methods, conjugate gradients. Applications.
Prerequisites
Basic knowledge of matrix algebra and real analysis. No specific background in differential geometry or numerical optimization will be assumed.
Evaluation procedure
The students will design and implement an algorithm for independent component
analysis and apply it to image processing or gene expression analysis problems.
The
task will be split into homeworks.
Schedule : 9h15 - 12h15
This course will take place at CESAME, Bâtiment Euler, 4, av. G. Lemaître, 1348 Louvain-la-Neuve
2. NUMERICAL
METHODS FOR NONLINEAR OPTIMAL CONTROL PROBLEMS
Lecturer : M. DIEHL (K.U.Leuven, Belgium)
The course develops numerical solution strategies for optimization
problems with underlying differential equation models. After a brief
overview over sequential and simultaneous approaches to optimal
control, we focus on the latter class, which solve both the model
equations and the optimality conditions in a "one-shot' approach. Here,
exploitation of sparsity in the linear solvers is of crucial
importance. Special topics include the treatment of distributed
inequalities by interior point methods, periodic problems. Finally, we
discuss real world applications from chemical and mechanical
engineering.
Contents
1. Optimal Control: Introduction and Overview
2. Dynamic Programming and Indirect Approaches (Pontryagin)
3. Direct Transcription Methods (ODE Sensititivity, Single Shooting)
4. Direct Collocation, Treatment of Sparsity
5. Direct Multiple Shooting
6. Applications and Extensions: Multi-Stage Processes, Periodicity
Prerequisites
Participants are assumed to have a knowledge of linear algebra and calculus. Basic knowledge in optimisation and numerical simulation of dynamic systems is advantageous.
Supporting material
Notes will be distributed during the course.
Related references
J. T. Betts: Practical Methods for Optimal Control Using Nonlinear Programming,
SIAM, Philadelphia, 2001
Bryson, A.E. and Ho, Y.-C.: Applied Optimal Control, Wiley, New York, 1975
Dates : November 26, 29 and December 03, 06, 07, 2007
Schedule : 9h15 - 12 h45
Location :
26 november : room 02.58
29 november, 03,06, 07 december : room 00.62
This
course will take place at the Katholieke Universiteit Leuven,
Department Elektrotechniek ESAT, Kasteelpark
Arenberg 1, 3001 Heverlee
SLIDES AND ARTICLES
http://homes.esat.kuleuven.be/~optec/events/courses/GraduateSchool/
3. MECHATRONICS : DYNAMICS OF ELECTROMECHANICAL AND PIEZOELECTRIC SYSTEMS,
WITH APPLICATIONS IN VIBRATION CONTROL
CANCELLED !!!
Contents
1) Lagrangian dynamics of mechanical systems :
Kinetic state functions, Virtual work principle, D'Alembert's principle, Hamilton's
principle, Lagrange's equation, gyroscopic systems, Jacobi integral, Rayleigh-Ritz
method.
2) Lagrangian dynamic of electrical network :
Constitutive equations of circuits elements, Hamilton's principle
(charge formulation and flux linkage formulation), Lagrange's equation (charge
formulation and flux linkage formulation)
3) Lagrangian dynamics of electromechanical systems :
Constitutive equations for transducers (movable-plate capacitor, movable-core
inductor, moving-coil transducer)
Hamilton's principle and Lagrange's equation (charge formulation
and flux linkage formulation)
Examples : loudspeaker, microphone, electrodynamic isolator, geophone, magnetic
suspension.
Control examples: sky-hook damper, magnetic levitation.
Self-sensing.
4) Piezoelectric systems :
Constitutive equations of piezoelectric transducers, electromechanical coupling
factor, structure with a piezoelectric transducer.
Piezoelectric material (constitutive equations, coenergy density function).
Hamilton's principle.
Rosen's transformer.
5) Piezoelectric laminates :
Beam actuator, Hamilton's principle, piezoelectric loads, laminar sensor, modal
actuator and sensor.
Beam with collocated actuator-sensor pair, pole-zero pattern, modal truncation.
Kirchhoff plate, multi-layer laminate.
6) Active and passive damping with piezoelectric transducers :
Active strut, Integral Force Feedback, voltage vs. current control, modal coordinates.
Damping via resistive shunting and inductive shunting.
Self-sensing.
4. ROBUST TRACKING AND DISTURBANCE REJECTION
Contents
Dates : April 15 ,16, 18, 21, 23, 25, 2008
Schedule : 9h15 - 12 h45
This course will take place at CESAME, Bâtiment Euler, 4, av. G. Lemaître, 1348 Louvain-la-Neuve
5. CONSTRUCTIVE LMI APPROACHES FOR ANTI-WINDUP COMPENSATOR DESIGN FOR SYSTEMS SUBJECT TO SATURATION
Physical, safety or technological constraints induce that the control actuators
can neither provide unlimited amplitude signals nor unlimited speed of reaction.
The control problems of combat aircraft prototypes and satellite launchers offer
interesting examples of the difficulties due to these major constraints. Neglecting
actuator saturations on both amplitude and dynamics can be source of undesirable
or even catastrophic behavior for the closed-loop system (such as loosing closed-loop
stability).
Such actuator saturations have also been blamed as one of several unfortunate
mishaps leading to the 1986 Chernobyl nuclear power plant disaster. For these
reasons, the study of the control problem (its structure, performance and stability
analysis) for systems subject to both amplitude and rate actuator saturations
as typical input constraints has received the attention of many researchers
in the last years.
Contents
1. General introduction
2. Context of stability
3. Approximating the saturation term
4. Introduction on anti-windup strategy
5. Solution via static anti-windup (1 loop)
6. Solution via static anti-windup (2 loops)
7. Solution via dynamical anti-windup
8. Concluding remarks and perspectives
Overview
The anti-windup approach consists of taking into account the effect of saturations in a second step after a previous design performed disregarding the saturation terms. The idea is then to introduce control modifications in order to recover, as much as possible, the performance induced by a previous design carried out on the basis of the unsaturated system. During this course, we will present some pertinent elements related to these anti-windup techniques and to illustrate their potentialities on some applications. In this context, LMI techniques have allowed to obtain constructive conditions to design the suitable anti-windup loops satisfying performance and stability requirements.
Prerequisites
Participants are assumed to have some background in linear algebra, Lyapunov theory, linear systems and control theory.
References
– J-M. Biannic, S. Tarbouriech, D. Farret, A practical approach
to performance analysis of saturated systems with application to
fighter aircraft flight
controllers, 5th IFAC Symposium ROCOND, Toulouse, France, July, 2006.
– Y.Y. Cao, Z. Lin, and D.G. Ward. An anti-windup approach to enlarging
domain of attraction for linear systems subject to actuator saturation.
IEEE
Transactions on Automatic Control, 47(1):140-145, 2002.
– J.M. Gomes da Silva Jr. and S. Tarbouriech, Anti-windup design with
guaranteed region of stability: an LMI-based approach, IEEE
Transactions on
Automatic Control, 50:1, pp.106-111, 2005.
– G. Grimm, J. Hatfield, I. Postlethwaite, A.R. Teel, M.C. Turner, and
L. Zaccarian. Anti-windup for stable linear systems with input
saturation: an LMI based synthesis. IEEE Transactions on Automatic
Control, 48(9):1509-1525, 2003.
– P. Hippe. Windup in control. Its effects and their prevention. AIC, Springer, Germany, 2006.
– S. Tarbouriech, G. Garcia, A.H. Glattfelder (Eds.), Advanced
strategies in control systems with input and output constraints, LNCIS,
vol.346, Springer Verlag, 2007.
– A.R. Teel, A nonlinear small gain theorem for the analysis of control
systems with saturation, IEEE Transactions on Automatatic Control, 41,
pp.1256-1270, 1996.
– M. Turner and L. Zaccarian (Editors). Special issue: Anti-windup.
International Journal of System Science, 37(2):65-139, 2006.
Dates : May 5, 6, 7, 13, 14, 15, 2008
Schedule : 9h15 - 12h15
This course will take place at CESAME, Bâtiment Euler, 4, av. G. Lemaître, 1348 Louvain-la-Neuve
Papers :
Registration
You can register electronically by filling in the following form via the web
:
http://www.inma.ucl.ac.be/graduate/graduate_registration.html
If you have problems with this, please contact Lydia DE BOECK
The admission is free for doctoral students and participants from Belgian academic
institutions. Other participants are requested to pay a registration fee of
EURO 500,- per course but a waiver can be obtained under special conditions
(contact the secretariat).
Payment can be made by bank transfer to the account n°310-0959001-48 with
the mention "AUTO2866 ACTIVITES DIDACTIQUES"
Secretariat